Advances in targeted therapy for malignant lymphoma

The incidence of lymphoma has gradually increased over previous decades, and it ranks among the ten most prevalent cancers worldwide. With the development of targeted therapeutic strategies, though a subset of lymphoma patients has become curable, the treatment of refractory and relapsed diseases remains challenging. Many efforts have been made to explore new targets and to develop corresponding therapies. In addition to novel antibodies targeting surface antigens and small molecular inhibitors targeting oncogenic signaling pathways and tumor suppressors, immune checkpoint inhibitors and chimeric antigen receptor T-cells have been rapidly developed to target the tumor microenvironment. Although these targeted agents have shown great success in treating lymphoma patients, adverse events should be noted. The selection of the most suitable candidates, optimal dosage, and effective combinations warrant further investigation. In this review, we systematically outlined the advances in targeted therapy for malignant lymphoma, providing a clinical rationale for mechanism-based lymphoma treatment in the era of precision medicine.


INTRODUCTION
Lymphoma is the most common lymphoid malignancy and is among the ten most prevalent cancers worldwide. 1 Lymphoma is a heterogeneous entity and includes Hodgkin's lymphoma (HL) and non-Hodgkin's lymphoma (NHL). HL accounts for 10-15% of lymphoma and is characterized by the presence of Reed-Sternberg cells. NHL accounts for 80-85% of lymphoma, including B-cell NHLs (B-NHLs) expressing CD20 or CD19, T-cell NHLs (T-NHLs) expressing CD3, CD4, or CD8, and natural killer (NK)/T-cell NHLs expressing CD56. Chemotherapy is the standard of care for lymphoma patients. The introduction of monoclonal antibodies targeting surface antigens has greatly changed the therapeutic landscape of lymphoma. For example, rituximab, an anti-CD20 antibody targeting CD20 in B-NHLs and brentuximab vedotin targeting CD30 in classical HL and T-NHLs, have significantly improved the response rates and clinical outcomes of patients. 2,3 In addition, growing insights into molecular biology and signaling pathways have led to the development of many innovative agents for lymphoma in recent years. 4 More recently, with a better understanding of the crosstalk between malignant lymphocytes and the tumor microenvironment, chimeric antigen receptor T-cells (CAR-T cells) have been rapidly developed in treating relapse and refractory patients. 5,6 Although the overall survival (OS) of lymphoma patients has been considerably improved by the new immunochemotherapeutic regimens, the selection of targeted agents and the optimal dosage are important due to treatmentrelated adverse events (AEs). In this review, we systematically outlined the advances in targeted therapy for malignant lymphoma that provide significant improvement in mechanism-based lymphoma treatment in the era of precision medicine.

SURFACE ANTIGENS AND TARGETED THERAPIES
Surface antigens are the most accessible part of lymphoma cells, and monoclonal antibodies (mAbs) targeting surface antigens have become important therapeutic strategies in many lymphoid malignancies. Cytotoxic to tumor cells, mAbs relatively spare normal tissues. The mechanisms of action include the induction of apoptosis, antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC). In addition to "bare" antibodies, antibodies or their fragments may be linked with cell toxins, immunotoxins, or radioisotopes to increase clinical efficacy.

CD20
The CD20 molecule is a transmembrane protein involved in B-cell activation and differentiation and is present on all mature B-cells and most B-NHL cells. 7 Moreover, without internalization or downregulation following antibody binding, CD20 functions as an ideal therapeutic target for most B-NHLs. 8 Moreover, pro-B cells and antibody-producing plasma cells do not express CD20, so anti-CD20 treatment will not impair the healthy B-cell population.
Anti-CD20 mAbs are classified as type I and type II. 9 Type I antibodies most effectively induce CDC, in which the binding of the mAb activates a complement cascade. Type I antibodies also induce ADCC, in which immune cells expressing Fc gamma receptor (FcγR) attack antibody-coated cells. Type II antibodies initiate ADCC as well as cell death through apoptotic or nonapoptotic mechanisms.
Rituximab was the first mAb to target CD20 and the first mAb approved to treat cancer patients. It is a chimeric antibody with a murine variable region and a human IgG1-kappa constant region, 8 classified as a type I mAb. The significant anti-lymphoma activity of rituximab in early trials 3,10-12 has led to its widespread use in most CD20 + B-NHLs.
The targeted agents and clinical trials related to mAbs are listed in Table 1. Ofatumumab is a fully humanized second-generation type I CD20 antibody that exhibits more potent CDC than rituximab in vitro. 13 Ofatumumab is approved in combination with chlorambucil for chronic lymphocytic leukemia (CLL). 14,15 Moreover, the results from a phase 2 trial (NCT00410163) suggested that ofatumumab in combination with fludarabine and cyclophosphamide was efficient in untreated CLL patients. 16 The main AEs were infusion-related reactions and grade 1-2 infections.
Obinutuzumab (GA101, Gazyva™) is a humanized type II mAb that can induce ADCC and direct apoptosis both in vitro and in vivo. 17,18 In a phase 1/2 study (NCT00517530), obinutuzumab as monotherapy showed clinical activity with an acceptable safety profile in aggressive B-NHLs. 19 Moreover, clinical trials (NCT01059630, NCT01332968, and NCT00825149) of obinutuzumab in combination with other chemotherapy regimens showed promising results in relapsed or refractory indolent B-NHLs 20,21 and untreated follicular lymphoma (FL). 22 The most common nonhematologic AEs were grade 1-2 infusion-related reactions, and the most common hematologic AE was neutropenia. For CLL, the findings of a phase 3 study (NCT01010061) of naïve elderly patients suggested that obinutuzumab in combination with chlorambucil yields better response rates and longer progression-free survival (PFS) than rituximab with chlorambucil and chlorambucil; thus, obinutuzumab became the first drug with "breakthrough therapy designation" approved by the FDA for the treatment of untreated CLL in combination with chlorambucil. 23 Recently, a multicenter, randomized, phase 3 trial (iLLUMINATE, NCT02264574) demonstrated the advantages of obinutuzumab plus ibrutinib over obinutuzumab plus chlorambucil as a first-line treatment for CLL. 24 Ublituximab is another type I, chimeric, recombinant IgG1 mAb targeting a unique epitope on the CD20 antigen, glycoengineered to enhance affinity for all FcRIIIa variants, leading to greater ADCC than other anti-CD20 mAbs such as rituximab and ofatumumab. 25 Ublituximab demonstrated efficacy and safety as a single agent in early clinical trials in patients with B-NHLs and CLL, 25,26 and it was further investigated in combination regimens. A phase 2 study (NCT02013128) combining ublituximab with ibrutinib was carried out in relapsed or refractory CLL and obtained an overall response rate (ORR) of 88%. Of note, in high-risk patients bearing del17p, del11q, or TP53 mutations, the ORR was 95%. 27 A phase 3 trial (GENUINE, NCT02301156) of ublituximab plus ibrutinib in high-risk relapsed or refractory CLL reported an ORR of 78% for the combination arm vs 45% for the monotherapy arm. 28 The combination of ublituximab and umbralisib with/without ibrutinib had indicated tolerability and activity in patients with relapsed or refractory B-NHLs and CLL in a phase 1 study (NCT02006485). 29,30 Other humanized type I anti-CD20 mAbs, such as veltuzumab (IMMU-106) and ocrelizumab (PRO70769), also showed efficacy in patients with relapsed or refractory B-NHLs and FL in phase 1/ 2 studies (NCT00285428 and NCT02723071). 31,32 In addition, progress has been made in the study of biosimilars of rituximab. CT-P10 (CELLTRION) was the first mAb biosimilar anticancer drug to gain international regulatory approval following the results of phase 3 trials (NCT02260804 and NCT02162771) in FL. 33,34 Other examples of rituximab biosimilars include GP2013, PF-05280586, and ABP798. GP2013 has also been approved in the European Union for its efficacy data from a phase 3 trial in FL (ASSIST-FL, NCT01419665). 35 The phase 3 study (NCT02213263) of PF-05280586 displayed positive results as well. 36 Moreover, ABP798 is currently under study (NCT02747043).
Radioimmunotherapy (RIT) has also emerged as an important therapeutic strategy for B-NHLs. Ibritumomab tiuxetan (IDEC-Y2B8, Zevalin®) is a radiolabeled anti-CD20 mAb that targets the same epitope on the CD20 molecule as rituximab. This compound chelates the radioactive particle yttrium-90 ( 90 Y), which delivers high beta energy to improve its ability to kill bulky, poorly vascularized tumors. 37 Ibritumomab tiuxetan is effective in both rituximab-naïve and rituximab-resistant FL, as well as in transformed B-NHLs. 38,39 Consequently, ibritumomab tiuxetan acquired FDA approval for rituximab-naïve relapsed or refractory low-grade B-NHLs and transformed NHLs. The long-term toxicity of developing myelodysplastic syndrome and acute myelogenous leukemia was observed. 40 Furthermore, ibritumomab tiuxetan has shown promising results in the first-line treatment of untreated FL (NCT00772655 and NCT01493479). 41,42 In addition, a phase 3 trial (FIT, NCT00185393) observed an improvement of efficacy through ibritumomab tiuxetan consolidation; 43,44 thus, the FDA approved this agent for consolidation therapy in untreated FL patients who achieve partial response (PR) or complete response (CR) after firstline chemotherapy. A phase 3 study of rituximab with or without ibritumomab tiuxetan in untreated FL is ongoing (NCT02320292). Ibritumomab tiuxetan is also being evaluated as consolidation therapy in relapsed or refractory FL in a phase 3 study (NCT01827605). Additionally, ibritumomab tiuxetan combined with high-dose chemotherapy prior to autologous stem cell transplantation (ASCT) has also been proven to be safe with relative efficacy. 45,46 CD22 CD22 is a single-spanning membrane glycoprotein with a molecular weight of 140,000 located on the surface of B-cells. It is mostly expressed in mature B-cells and many malignant Bcells. 47,48 CD22 acts as a negative regulator of B-cell receptor (BCR)-induced signaling and plays a critical role in B-cell activation. 47,49 The inhibitory function of CD22 and its restricted expression on B-cells make CD22 an ideal target in NHLs.
Epratuzumab is a humanized IgG1 mAb targeting CD22. The crosslinking of CD22 by epratuzumab triggers BCR signaling and caspase-dependent apoptosis in human lymphoma cells. 50 Preclinical studies demonstrated that CD22 mAbs had independent lymphomacidal properties. 51 Single-agent epratuzumab has been investigated in both indolent and aggressive NHLs. In an early phase 1/2 trial including 55 patients with recurrent NHLs, epratuzumab showed a response in FL (ORR 24%), while no response was observed in other indolent lymphomas. 52 In another concurrent phase 1/2 trial, 15% of patients with relapsed or refractory diffuse large B-cell lymphoma (DLBCL) responded to epratuzumab. 53 The combination of epratuzumab with rituximab has been tested in a multicenter phase 2 trial and exhibited an ORR of 54% in FL and 57% in small lymphocytic lymphoma (SLL). 54 Epratuzumab plus rituximab was also studied in untreated FL and obtained an ORR of 88.2% (NCT00553501). 55 In aggressive lymphomas, a phase 2 trial (NCT00301821) showed that epratuzumab combined with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP) achieved an ORR of 96% in DLBCL, with 3-year event-free survival (EFS) and OS rates of 70% and 80%, respectively. 56 Conjugate antibodies utilize the direct conjugation of mAbs with cytotoxic agents, and there are two types of antibody-based conjugates: antibody-drug conjugates (ADCs) and immunotoxins. 57 ADCs are mAbs connected to bioactive drugs by chemical linkers. Inotuzumab ozogamicin (InO, CMC-544) is a CD22-targeted ADC combining a humanized IgG4 anti-CD22 mAb with calicheamicin, an enediyne antibiotic, which causes DNA damage and cell apoptosis. 58,59 The combination of InO with rituximab in a phase 1/2 study (NCT00299494) of relapsed FL, DLBCL, and refractory aggressive NHL induced ORRs of 87%, 74%, and 20%, respectively. The most common grade 3-4 AEs were thrombocytopenia (31%) and neutropenia (22%). 60 However, InO plus rituximab failed to obtain positive results in a randomized phase 3 trial (NCT01232556) of relapsed or refractory CD22 + aggressive B-NHLs and FLs. 61 64 Moreover, a pivotal phase 3 study (NCT01829711) for relapsed or refractory HCL obtained an ORR of 75%, with a CR rate of 41%. 65 The FDA approved moxetumomab pasudotox (Lumoxiti) for the treatment of adult patients with relapsed or refractory HCL.

CD30
CD30 is a 120-kDa type I transmembrane receptor of the tumor necrosis factor receptor (TNFR) superfamily. 66 The binding of CD30 with its ligand induces signal transduction through several downstream pathways, especially nuclear factor-κB (NF-κB). 67 CD30 is normally expressed on activated B cells, T cells, and NK cells, as well as virally infected lymphocytes. In addition, CD30 is universally expressed in HL and anaplastic large cell lymphoma (ALCL). 68,69 Other lymphoproliferative disorders, such as DLBCL, primary mediastinal B-cell lymphoma (PMBCL), peripheral T-cell lymphoma (PTCL), mycosis fungoides (MF), Sézary syndrome (SS) and adult T-cell leukemia/lymphoma (ATLL), can also express CD30 to various degrees. [70][71][72] A chimeric mAb SGN-30, consisting of the variable region of an anti-CD30 murine mAb with human gamma 1 heavy chain and kappa light chain constant regions, promotes growth arrest and DNA fragmentation in vitro and exhibits antitumor activity in HL models. 73 In a phase 2 study of relapsed or refractory HL or ALCL, SGN-30 showed only a modest effect in ALCL (2 CR and 5 PR in 41 ALCL patients). 74 However, another phase 2 trial used a combination of SGN-30 with gemcitabine, vinorelbine, and liposomal doxorubicin in relapsed HL and showed an ORR of 65%, while grades 3-5 pneumonitis occurred in five patients, leading to the premature closure of the trial. 75 Brentuximab vedotin (BV, Adcetris), a CD30 ADC, connects an anti-CD30 antibody with the anti-mitotic agent monomethyl auristatin E (MMAE) via a valine-citrulline peptide-linker. It showed strong activity against CD30 + tumor cell lines in vitro, as well as xenograft models of HL and ALCL. 76 A phase 1 dose-escalation study (NCT00430846) of BV in 45 patients with relapsed or refractory CD30 + hematological malignancies (mainly HL) determined the optimal dose of BV as 1.8 mg/m 2 intravenously every 3 weeks and showed an ORR of 38%. 2 Common AEs of BV include fatigue, pyrexia, diarrhea, nausea, peripheral neuropathy, neutropenia, anemia, and arthralgias. 2 Other AEs, such as anaphylaxis and acute pancreatitis, have also been reported. 77,78 BV was granted FDA accelerated approval for the treatment of relapsed or refractory HL and ALCL based on the results of two phase 2 studies. NCT00848926 enrolled 102 relapsed or refractory HL patients and obtained an ORR of 75% (CR 34%) with a median duration of response (DoR) of 6.7 months. 79 NCT00866047 showed an ORR of 86% (CR 57%) with a median DoR of 12.6 months in 58 patients with relapsed or refractory CD30 + ALCL. 80 After approval, the FDA issued a boxed warning related to the risk of progressive multifocal leukoencephalopathy and added a contraindication warning for the concomitant use of BV and bleomycin due to pulmonary toxicity.
In addition to ALCL, BV has shown efficacy as a single agent in other T-NHLs (NCT01421667). 81 In addition to systemic lymphomas, BV was also utilized in primary CD30 + cutaneous lymphomas and showed encouraging efficacy. 82 A phase 3 randomized multicenter trial (ALCANZA, NCT01578499) was conducted to evaluate single-agent BV vs a control arm of the investigator's choice of standard therapies in patients with CD30 + primary For BV combined with chemotherapy, in a multicenter phase 3 trial (NCT01712490) involving patients with untreated stage III or IV HL, patients were randomized to receive BV, doxorubicin, vinblastine, and dacarbazine (A+AVD) or doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD). The results showed that at a median follow-up of 24.6 months, the 2-year modified PFS rates in the A+AVD and ABVD groups were 82.1% and 77.2%, respectively. Neutropenia and peripheral neuropathy were the most common AEs. 84 Based on these promising clinical data, the FDA expanded the approval of BV for the first-line treatment of stage III or IV HL in combination with chemotherapy. A phase 1 study (NCT01309789) combining BV with cyclophosphamide, doxorubicin, and prednisolone in patients with CD30 + PTCL resulted in an objective response in all patients (CR 88%). 85 Moreover, the five-year followup demonstrated durable remission in half of the patients after combination therapy. 86 Therefore, a randomized phase 3 trial (ECHELON-2, NCT01777152) comparing BV plus cyclophosphamide, doxorubicin and prednisone (CHP) with CHOP was conducted in untreated patients and demonstrated a significant improvement in PFS and OS with a manageable safety profile when using BV plus CHP. 87 The FDA thus approved BV in combination with chemotherapy for adults with untreated ALCL or other CD30 + PTCL.

CD52
The CD52 antigen is a small glycopeptide highly expressed on normal and malignant B and T lymphocytes. The exact function of CD52 remains undefined, but in vitro studies have proven that it is a costimulatory molecule for the induction of CD4 + regulatory T-cells. 88 Alemtuzumab (Campath®) is a humanized mAb targeting CD52 that can induce complement-mediated lysis as well as caspaseindependent cell death in malignant lymphoid cells. 89,90 Singleagent alemtuzumab received accelerated approval by the FDA for CLL patients who had received alkylating agents and failed fludarabine therapy. 91 A phase 3 randomized trial comparing alemtuzumab to chlorambucil as first-line treatment showed significantly improved PFS, time to alternative treatment, ORR and CR, with manageable toxicity in CLL. 92 Alemtuzumab has also been evaluated as monotherapy in T-NHLs and exhibited efficacy in advanced MF, Sézary syndrome (SS), and relapsed or refractory PTCL, 93,94 where hematological toxicity and cytomegalovirus (CMV) reactivation were the most common AEs.
Alemtuzumab-containing chemoimmunotherapy regimens can be effective but have been limited by their toxicities in CLL (NCT00564512). 95 The bendamustine and subcutaneous alemtuzumab combination was proven to be as effective as the combination of fludarabine, cyclophosphamide, and cladribine and was safe in heavily pretreated and elderly patients. 96 Other attempts at combining pentostatin, alemtuzumab, and low-dose rituximab (NCT00669318) also yielded efficacy and tolerability in relapsed or refractory 17p13-deleted CLL. 97 The combination of alemtuzumab and CHOP-based chemotherapy was explored in untreated PTCL. 98-100 Phase 3 randomized studies (NCT00646854 and NCT00725231) of alemtuzumab plus CHOP in either young or elderly PTCL patients achieved improved PFS or OS. 101,102 CD79 CD79, composed of CD79A and CD79B components, is a main BCR signaling component and is expressed almost exclusively on B-cells and B-NHLs. CD79 expression precedes immunoglobulin heavy-chain gene rearrangement and CD20 expression during B-cell development but disappears in the late stage of B-cell differentiation. 103 When BCR is cross-linked, CD79 is targeted to a lysosome-like compartment 104 and induces cell apoptosis or triggers cell activation and division with rescue signals from T cells. 105 Therefore, CD79 has become an attractive target for the use of ADCs, and preclinical studies found two stable-linker ADCs capable of killing NHL cell lines in vitro and in xenograft models. 106 Polatuzumab vedotin (DCDS4501A) is an anti-CD79B mAb conjugated to MMAE. In a phase 1 study (NCT01290549) in relapsed or refractory B-NHLs and CLL, no objective response was observed in CLL, while at the recommended phase 2 dose of 2.4 mg/kg, objective responses were obtained in 23 of 42 patients with NHLs by polatuzumab vedotin monotherapy (56% in patients with DLBCL, 47% with indolent NHLs, and 100% with mantle cell lymphoma (MCL)) and in 7 of 9 patients by polatuzumab vedotin plus rituximab. 106 Polatuzumab vedotin was further evaluated in a phase 2 trial (NCT01691898) in combination with rituximab in patients with relapsed or refractory NHLs. The results showed that the ORRs and CR rates were 54% and 21% in DLBCL and 70% and 45% in FL, respectively. Grade ≥3 AEs occurred in 77% of DLBCL patients and 50% of FL patients, mainly as neutropenia, anemia, and diarrhea. 107 Furthermore, the findings of a phase 2 study (NCT02257567) pointed out that adding polatuzumab vedotin to bendamustine and rituximab (BR) treatment improved survival in patients with relapsed or refractory DLBCL. 108 The combination of polatuzumab vedotin with rituximab, cyclophosphamide, doxorubicin and prednisone (R-CHP) vs R-CHOP in DLBCL is currently being investigated in a phase 3 study (POLARIX, NCT03274492).

CD19
CD19 is a B-cell-specific member of the immunoglobulin superfamily that augments signals by the pre-BCR/BCR and modulates B-cell fate decisions at multiple stages of development. 109 CD19 is highly expressed in nearly all B-NHLs, making it an excellent target for immune-based therapies. 110 Inebilizumab (MEDI-551) is a CD19-targeted humanized mAb that has potent ADCC activity in vitro and in vivo in preclinical studies. 111 Inebilizumab monotherapy has been evaluated in phase 1 studies and showed acceptable toxicity and promising efficacy in patients with relapsed or refractory FL and DLBCL (NCT01957579). 112 A phase 1/2 trial (NCT00983619) of inebilizumab alone and in combination with rituximab in FL, CLL, and DLBCL has recently been completed. Regarding inebilizumab in combination with chemotherapy, recent clinical trials did not yield promising results. A phase 2 trial (NCT01466153) comparing inebilizumab plus bendamustine and BR did not find any significant difference in the ORR between the two groups. Another randomized phase 2 study (NCT01453205) on rituximab plus ifosfamide, carboplatin, and etoposide (ICE)/dexamethasone, high-dose cytarabine, and cisplatin (DHAP) vs inebilizumab plus ICE/DHAP in patients with relapsed or refractory DLBCL did not show any significant difference in ORR, PFS, or OS.
Tafasitamab (MOR208, XmAb®5574) is a novel Fc-engineered, humanized, anti-CD19 antibody with enhanced ADCC, antibodydependent cellular phagocytosis and apoptosis, as well as more potent antitumor activity in vivo than its IgG1 analog. 113 These effects were achieved by increasing the affinity for FcγRIIIa on effector cells through the introduction of S239D and I332E amino acid substitutions to the Fc domain. Tafasitamab monotherapy exhibited promising clinical activity in patients with relapsed or refractory B-NHLs with a favorable safety profile. The ORRs were 26%, 29%, and 27% in DLBCL, FL, and other indolent NHLs, respectively, with 9% of patients experiencing grade 3-4 neutropenia (NCT01685008). 114 Furthermore, combinations with lenalidomide and bendamustine are being evaluated in recent phase 2/3 clinical trials (NCT02399085, NCT02005289, and NCT02763319). Based on the preliminary data from a phase 2 study (L-MIND, NCT02399085) in combination with lenalidomide, this mAb was granted FDA breakthrough therapy and fast track designations for DLBCL. Eighty-one patients enrolled in the L-MIND study obtained an ORR of 58%, including 33% CR, with no unexpected toxicities observed. With a median follow-up of 12 months, the median PFS was 16.2 months. 115 In addition, the CD19-targeted ADC coltuximab ravtansine (SAR3419) consists of a cytotoxic maytansinoid, DM4, which is a potent inhibitor of tubulin polymerization and microtubule assembly. In a phase 2 study (NCT01472887), this agent showed good tolerance and moderate clinical responses in pretreated patients with relapsed or refractory DLBCL (ORR 43.9%). 116 A novel ADC based on coltuximab ravtansine showed promising preclinical data and may become an attractive candidate for clinical investigation. 117 Loncastuximab tesirine (ADCT-402) is a novel CD19-targeted ADC that delivers SG3199, a highly cytotoxic pyrrolobenzodiazepine dimer, and showed highly targeted cytotoxicity in vitro and antitumor activity in vivo in preclinical studies. 118 A pivotal phase 2 study (NCT03589469) is currently ongoing on relapsed or refractory DLBCL, as well as phase 1 studies (NCT02669017, NCT03684694, and NCT03685344) on relapsed or refractory B-NHLs.

CD37
CD37 is a heavily glycosylated transmembrane protein of the tetraspanin superfamily and represents one of the specific proteins for normal and malignant mature B-cells. The expression of CD37 is detected in CLL, Burkitt lymphoma (BL), MCL, and FL, 119,120 and it is involved in various biological processes, such as cell adhesion, proliferation, differentiation, intercellular communication via exosomes and immune response. 121 Small modular immunopharmaceuticals (SMIPs) are disulfidelinked single-chain proteins comprised of one antigen-binding region (V H /V L ), a hinge, and an Fc domain of the human IgG1 region (CH2-CH3). Due to their smaller size, SMIPs may have better tissue penetration than mAbs. SMIP-016 is a homodimeric protein specially engineered to exhibit the full binding activity of an anti-CD37 antibody. Preclinical studies have demonstrated that SMIP-016 can induce apoptosis and ADCC in B-cell leukemia/lymphoma cell lines and primary CLL cells. 122 Otlertuzumab (TRU-016) is a humanized variant of SMIP-016 built on the ADAPTIR (modular protein technology) platform. In a phase 1 study (NCT00614042), otlertuzumab was well tolerated and exhibited modest activity as monotherapy in CLL and select subtypes of relapsed or refractory NHLs. The ORR was 23% in CLL, with the most frequent grade ≥3 AEs being thrombocytopenia, neutropenia, anemia, fatigue, and hypophosphatemia. 123 For patients with relapsed or refractory FL, MCL, and Waldenström's macroglobulinemia (WM), a lymph node reduction of 50% or more was observed in 3 of 12 patients. 124 The efficacy of this agent can be enhanced in combination with chemotherapy. A randomized phase 2 trial (NCT01188681) showed a significantly increased response rate and prolonged PFS of otlertuzumab in combination with bendamustine over single-agent bendamustine in relapsed CLL. The ORR of this combination therapy was 69%, with a median PFS of 15.9 months. 125 Similarly, a phase 1 study (NCT01317901) combining otlertuzumab with BR in relapsed or refractory B-NHLs showed promising activity with no unexpected toxicity. The ORR was 83% (CR 32%). 126 Anti-CD37 ADCs such as IMGN529 and AGS67E were also studied. IMGN529 couples an anti-CD37 antibody with the maytansinederived anti-microtubule agent, DM1. IMGN529 has exhibited potent antitumor activity in preclinical models of CD37 + NHLs. 127,128 A phase 1 trial (NCT01534715) of IMGN529 in relapsed or refractory NHLs and CLL has recently been reported, showing manageable safety profiles and preliminary evidence of activity, particularly in DLBCL. 129 AGS67E is a fully human monoclonal IgG2 antibody conjugated via a protease-cleavable linker to MMAE. AGS67E has shown remarkable preclinical antitumor effects in NHLs and CLL cell lines and patient-derived xenograft models. 130 Clinically, a phase 1 study (NCT02175433) of escalating doses of AGS67E as monotherapy in relapsed or refractory lymphoid malignancies is ongoing. 177 Lu-lilotomab satetraxetan ( 177 Lu-DOTA-HH1, Betalutin®) is a novel antibody radionuclide conjugate (ARC) targeting the CD37 antigen. This agent received fast channel assignment from the FDA based on the preliminary data of efficacy and safety in a phase 1/2 trial (LYMRIT 37-01, NCT01796171) in relapsed or refractory FL. It is currently in a pivotal phase 2 trial (PARADIGME) in third-line rituximab-resistant FL, while also being investigated as a single agent in a phase 1 study (NCT02658968) in relapsed or refractory DLBCL and in combination with rituximab in a phase 1 study (NCT03806179) in second-line FL treatment.
C-C chemokine receptor type 4 C-C chemokine receptor type 4 (CCR4) is a seven-transmembrane G-protein-coupled receptor principally expressed on Th2 cells and CD4 + regulatory T cells, 131,132 as well as in various types of PTCLs, including MF and ATLL. 133,134 Furthermore, CCR4 expression was found to be an independent and significant unfavorable prognostic factor in these diseases, 133,134 which makes it a promising target in the treatment of PTCL and ATLL.
Mogamulizumab (KW-0761, Poteligeo) is the first defucosylated humanized mAb directed against CCR4; it has been proven to induce ADCC against CCR4 + malignant T cells 135 and to reduce CCR4 + Treg cell numbers in cutaneous T-cell lymphoma (CTCL). 136,137 Mogamulizumab was first approved for relapsed or refractory ATLL due to its promising efficacy (ORR 50%) and acceptable toxicities in a phase 2 study (NCT00920790). 138 In a randomized phase 2 study (NCT01173887) of dose-intensified chemotherapy with or without mogamulizumab in untreated aggressive ATLL, the mogamulizumab-containing arm showed a higher CR rate with manageable toxicities. 139 In addition to its application in ATLL, the efficacy of mogamulizumab in CTCL has also been confirmed. A phase 1/2 study (NCT00888927) of mogamulizumab was performed on 41 pretreated patients with CTCL and resulted in an ORR of 36.8% (47.1% in SS and 28.6% in MF). The most common AEs were nausea, chills, and infusionrelated reactions. 140 A multicenter phase 2 study (NCT01192984) of relapsed CCR4 + PTCL and CTCL patients in Japan obtained an ORR of 35% and a median PFS of 3 months. Lymphocytopenia, leukocytopenia, and neutropenia (19%) were the most common grade 3-4 AEs. 141 Therefore, mogamulizumab was first approved for untreated ATLL as well as relapsed or refractory PTCL in Japan.
The final results of a phase 3, randomized, multicenter clinical trial of mogamulizumab vs vorinostat in previously treated CTCL (MAVORIC, NCT01728805) have been reported. 142 The study included 372 patients and was the largest randomized trial in CTCL. Mogamulizumab resulted in a longer PFS than vorinostat (median 7.7 months vs. 3.1 months). The most common AEs of mogamulizumab were pyrexia and cellulitis. Mogamulizumab was granted approval in the European Union and the United States for the treatment of adult patients with relapsed or refractory MF or SS after at least one prior systemic therapy. 143 Other surface antigens CD25. CD25 (IL2R-α) is expressed on both HL and various NHLs and has been studied as a therapeutic target for over two decades. Denileukin diftitox (DD, ONTAK), a diphtheria exotoxin conjugated to an IL-2 fragment, was granted full FDA approval for the treatment of CTCL. 144 Although the efficacy of the anti-CD25 antibodies basiliximab and daclizumab is limited, radiolabeled antibodies are promising. 90 Y-daclizumab achieved responses in 50% of patients with relapsed HL (NCT00001575). 145 90 Ybasiliximab is being evaluated in combination with carmustine, etoposide, cytarabine, melphalan (BEAM) chemotherapy for ASCT in relapsed or refractory HL (NCT01476839), as well as T-NHLs (NCT02342782). Camidanlumab tesirine (ADCT-301), a CD25 ADC, has been investigated in a phase 1 trial (NCT02432235) in patients with CD25 + relapsed or refractory HL and NHLs.
CD38. The CD38 antigen is a type II transmembrane glycoprotein with receptor and enzyme functions that is expressed in a number of hematological malignancies, particularly in multiple myeloma (MM). 146 In addition, its expression has also been reported in lymphomas such as MCL 147 and NK/T-cell lymphoma (NKTCL). 148 Daratumumab is a CD38 mAb approved for treating relapsed or refractory and untreated MM. In a phase 2 study (NCT02927925) of daratumumab in relapsed or refractory NKTCL, the ORR was 35.7% in 16 patients. 149 CD40. CD40 is a type-I transmembrane protein that belongs to the TNFR family. CD40 is expressed on B cells, monocytes, dendritic cells, endothelial cells and epithelial cells and plays a critical role in the regulation of immune responses. 150 In addition, CD40 is expressed on B-NHLs, leading to the modulation of tumor cell growth after binding with its natural ligand (CD40L). 151 Dacetuzumab (SGN-40) is a humanized IgG1 mAb targeting CD40. Although dacetuzumab has previously demonstrated antilymphoma activity in a phase 1 study (NCT00103779), 152 singleagent dacetuzumab showed only modest activity in patients with relapsed DLBCL (NCT00435916) 153 and failed to obtain higher CR rates when combined with rituximab plus ICE (R-ICE) in relapsed DLBCL in a phase 2 study (NCT00529503). 154 CD74. The humanized antibody milatuzumab (hLL1) is a mAb against CD74, which is involved in malignant B-cell proliferation and survival. Preclinical studies found that milatuzumab had promising antitumor activity in NHL in vitro and in tumor xenograft models. 155 Moreover, a phase 1/2 study (NCT00989586) delivered the anti-CD20 mAb veltuzumab (200 mg/m 2 weekly) and escalating doses of milatuzumab to relapsed or refractory B-NHL patients and reported an ORR of 24% and a median DoR of 12 months. 156 Another preclinical study of the novel bispecific hexavalent Abs (HexAbs) veltuzumab and milatuzumab demonstrated enhanced antitumor activity in cell lines or primary patient samples of MCL and other CD20 + /CD74 + malignancies. 157 CD80. CD80 (B7-1), a cell-surface receptor, is implicated in the costimulation of T-cell function and expressed on B-NHLs. The anti-CD80 mAb galiximab (IDEC-114) can inhibit tumor cells of B-NHLs in vitro and in mouse models, either alone or combined with chemotherapy (fludarabine or doxorubicin). 158 A phase 2 study (NCT00516217) evaluated galiximab in relapsed or refractory HL and reported an ORR of 10.3%. Moreover, a phase 1/2 study on galiximab in relapsed or refractory FL revealed an ORR of 11% (CR 6%). 159 Another phase 1/2 trial (NCT00048555) of galiximab and rituximab reported an ORR of 66% (CR 19% and unconfirmed complete remission (CRu) 14%) in relapsed or refractory FL with rituximab-refractory patients excluded. 160 CD158k. CD158k (KIR3DL2) is a member of the highly polymorphic family of killer-cell immunoglobulin-like receptors (KIRs) and is expressed on NK cells and a small proportion of CD8 + T cells, as well as CD4 + T cells in CTCL. [161][162][163] The anti-CD158k mAb IPH4102 has been found to be potent and safe in preclinical studies. 164 A phase 1 study (NCT02593045) demonstrated efficacy and safety in CTCL, 165 with the expansion study ongoing. In addition, a phase 2 study (NCT03902184) of IPH4102 alone or in combination with chemotherapy is recruiting patients with advanced T-NHLs.
Bispecific T cell Engagers. Bispecific T cell Engagers (BiTEs) are engineered bispecific anti-CD3 antibodies consisting of the variable domains of two antibodies linked in a single chain. A BiTE antibody binds both CD3 + cytotoxic T cells and a target antigen to bring the two cells into proximity and thus triggers T cells to kill tumor cells via perforin-mediated apoptosis. 166 Blinatumomab is a CD19/CD3 BiTE that shows remarkable antilymphoma activity both in vitro and in vivo. 167,168 In a phase 1 dose-escalation study (NCT00274742) in patients with relapsed or refractory NHLs, 60 μg/m 2 /day was established as the maximum tolerated dose, with 22% of patients experiencing grade 3 neurologic events. For patients treated at 60 μg/m 2 /day, the ORR was 69% (DLBCL, 55%; MCL, 71%; FL, 80%), with a median DoR of 404 days. 169 In another phase 2 study (NCT01741792) in patients with relapsed or refractory DLBCL comparing weekly step-up dosing with flat dosing, the ORR was 43%. However, neurological AEs are also common. 170 A later phase 2 trial (NCT02910063) of blinatumomab in aggressive B-NHLs is ongoing.
In summary, therapies targeting the lymphoma surface antigen have made great progress. In general, mAbs are effective in the treatment of lymphoma, as evidenced by the FDA accelerated approval of many drugs. Moreover, mAbs as monotherapy have fewer adverse reactions and higher tolerance than conventional chemotherapy. However, mAbs also have limitations, such as offtarget effects. In the future, more research on the precise mechanisms of the efficacy and resistance of mAbs is needed. The design of future clinical trials should focus on subgroups with specific pathogenic mechanisms. At the same time, attention should also be paid to the timing, duration, and dose optimization of mAbs, either alone or in combination with traditional chemotherapy.

SIGNALING TRANSDUCTION PATHWAYS AND TARGETED THERAPIES
Signaling transduction pathways are critically involved in lymphoma progression. Inhibitors targeting key pathways, including spleen tyrosine kinase (SYK), Bruton's tyrosine kinase (BTK), phosphoinositide 3-kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR), Janus kinase-signal transducer and activator of transcription (JAK-STAT), NOTCH, NF-κB and ubiquitin-proteasome pathway (UPP), have been applied to treat lymphomas. SYK SYK, a nonreceptor tyrosine kinase, plays an important role in BCR and T-cell receptor (TCR) signaling. The phosphorylation of immunoreceptor tyrosine-based activation motifs (ITAMs) in the Igα (CD79A)/Igβ (CD79B) cytoplasm region recruits SYK and induces SYK activation, BTK recruitment, and phospholipase Cγ2 (PLCγ2) activation. 171 In TCR signaling, phosphorylated CD3 and ζ subunits of the TCR complex by the Src-related kinases LCK and FYN recruit zeta-chain-associated protein kinase 70 (ZAP-70) and SYK (Fig. 1). 172 The activated B cell-like subtype of DLBCL (ABC-DLBCL) is characterized by antigen-driven BCR signaling, 173,174 while germinal center B cell-like (GCB)-DLBCL features tonic, antigenindependent BCR signaling. 175,176 BL is also characterized by tonic BCR signaling and mostly relies on SYK. 177 In T-NHLs, aberrant SYK expression was reported in monomorphic epitheliotropic intestinal T-cell lymphomas (MEITL, type II EATL), 178 the follicular variant of PTCL, not otherwise specified (PTCL-NOS), and angioimmunoblastic T-cell lymphoma (AITL) due to t(5;9)(q33;q22) ITK/SYK translocation. [179][180][181] The targeted agents and clinical trials related to SYK and BTK are listed in Table 2. Fostamatinib disodium, the first approved oral SYK inhibitor, was evaluated in a phase 1/2 trial (NCT00446095) of recurrent B-NHLs, showing an ORR of 22% in DLBCL, 10% in FL, and 11% in MCL. 182 TAK-659 is being studied in a phase 2 trial in relapsed or refractory DLBCL (NCT03123393) alone, in combination with venetoclax in NHLs in a phase 1 trial (NCT03357627), and in combination with R-CHOP in DLBCL in a phase 1 trial (NCT03742258). The efficacy of entospletinib (GS-9973) is being explored in a phase 2 trial (NCT01799889) in relapsed or refractory hematologic malignancies alone as well as in combination with obinutuzumab in a phase 1/2 trial in NHLs (NCT03010358). Another phase 2 study (NCT01796470) of entospletinib combined with idelalisib in relapsed or refractory NHLs and CLL underwent early termination due to treatment-emergent pneumonitis in 18% of patients. 183 Cerdulatinib (PRT-062070), a dual SYK/JAK inhibitor, was reported to have a greater capacity to suppress cell proliferation and induce apoptosis than PRT-060318, an SYKselective inhibitor, in ATLL-derived cell lines and murine models. 184 A phase 1/2 trial (NCT01994382) of cerdulatinib in NHLs and CLL/SLL and a phase 2 trial (NCT04021082) of cerdulatinib in relapsed or refractory PTCL are ongoing.

BCR-BTK
The activation of BCR leads to the phosphorylation of LYN and SYK, which phosphorylate tyrosine residues in the cytoplasmic part of CD19 and B-cell adaptor for PI3K (BCAP), inducing PI3K activation, phosphatidylinositol 4,5-bisphosphate (PIP2) transformation to phosphatidylinositol 3,4,5-trisphosphate (PIP3), and BTK recruitment. BTK activation leads to PLCγ2 phosphorylation, which could further hydrolyze PIP2 to produce 1,4,5-trisphosphate (IP3) and diacylglycerol (DAG). 185 IP3 is involved in intracellular calcium regulation and nuclear factor of activated T cells (NFAT) transcription, and DAG is associated with protein kinase Cβ (PKCβ) and mitogen-activated protein kinase (MAPK) family activation. 186 PKCβ also participates in the NF-κB pathway through a scaffold complex including CARD11, BCL-10, and MALT1. BTK plays a key role in the tonic BCR signaling pathway through the positive regulation of AKT phosphorylation (Fig. 1). 187 The inhibition of BTK decreased BTK Y223 phosphorylation and anti-apoptotic protein expression (BCL-2, BCL-XL, and MCL-1), resulting in increased apoptosis in MCL cell lines. 188 Moreover, recurrent gene mutations of the BCR-BTK signaling pathway are frequently found in ABC-DLBCL, FL, and marginal zone lymphoma (MZL). 4,[189][190][191][192] Ibrutinib is an irreversible BTK inhibitor that suppresses BTK enzymatic activity through a covalent bond with a cysteine residue in the BTK active site. A phase 1/2 study (NCT00849654) of ibrutinib enrolled patients with relapsed or refractory B-NHLs and reported promising safety and response (ORR 60% and CR 16%). 193 In a phase 1/2 trial (NCT01325701) of relapsed or refractory DLBCL, ibrutinib induced an ORR of 37% in ABC-DLBCL but only an ORR of 5% in GCB-DLBCL. 194 A phase 2 trial (NCT01849263) of ibrutinib in relapsed or refractory FL reported an ORR of 37.5% (CR 12.5%). 195 Ibrutinib has also been actively investigated in other relapsed or refractory B-NHLs and has shown clinical efficacy (NCT01980628 and NCT01236391). 196,197 A phase 1/2 trial (NCT02329847) of ibrutinib in combination with nivolumab in relapsed or refractory B-cell malignancies revealed an ORR of 36% in DLBCL (CR 16%), 33% in FL (CR 10%), and 61% in CLL/SLL (CR 0%). 198 Moreover, a phase 2 study (NCT02471391) of ibrutinib combined with venetoclax in MCL reported an ORR of 71% (CR 62%). 199 The combination of ibrutinib, lenalidomide, and rituximab is being explored in a phase 2 trial (NCT03949062) to evaluate its efficacy and safety in untreated and unfit elderly DLBCL patients. This combination also induced an ORR of 95% in untreated FL in a phase 1 trial (NCT01829568), as well as an ORR of 76% (CR 56%) in relapsed or refractory MCL in a phase 2 trial (NCT02460276). 200,201 In untreated CD20 + B-NHLs, ibrutinib plus R-CHOP achieved an ORR of 100% in a phase 1 study (NCT01569750). 202 In addition, in a phase 3 study (NCT01855750) in untreated non-GCB DLBCL, ibrutinib plus R-CHOP produced a CR rate of 67.3%, and placebo plus R-CHOP produced a CR rate of 68.0%, with no statistically significant difference. Moreover, the sequential combination of ibrutinib with high-dose methotrexate and rituximab was studied in patients with primary central nervous system lymphoma (PCNSL) (NCT02315326). 203 Acalabrutinib (ACP-196) is a BTK inhibitor that has been proven to have a more enhanced efficacy than ibrutinib in canine studies. 204 A phase 2 study (NCT02213926) reported an ORR of 81% (CR 40%) in relapsed or refractory MCL. 205 The FDA has approved acalabrutinib for treating relapsed or refractory MCL. Moreover, in a phase 1/2 trial (NCT02029443) of acalabrutinib in relapsed CLL, the ORR was 95%, and a 100% ORR was

PI3K-AKT-mTOR
The PI3K-AKT-mTOR pathway is an important regulator in normal myeloid and lymphoid development. 210 Upon activation, BCAP is upregulated, and the catalytic subunit of PI3K (referred to as p110α, p110β, p110γ, and p110δ for the four different isoforms) triggers PIP3 and recruits the serine/threonine kinase AKT to the plasma membrane. 211 AKT can subsequently activate mTOR, which encompasses two different multiprotein complexes, mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2). mTORC1 phosphorylates 4E-BP1 and S6K1 to activate key drivers of protein translation (Fig. 1). 212 Downstream signaling of BCR is largely dependent on p110δ, and mutations in PIK3CA (the gene encoding p110α) were found in approximately 1-8% of DLBCL. 213,214 In T-NHLs, p110δ and p110γ are vital kinases of TCR signaling and chemokine receptor signaling, respectively. [215][216][217] The targeted drugs and clinical trials related to the PI3K-AKT-mTOR, JAK-STAT, NOTCH, and NF-κB signaling pathways are listed in Table 3. Idelalisib (CAL-101, GS-1101), a p110δ-selective inhibitor, is the first FDA-approved PI3K inhibitor in treating relapsed FL and SLL. A phase 2 trial (NCT01393106) demonstrated that idelalisib was tolerable and had modest single-agent activity in relapsed or refractory HL (ORR 68% and CR 4%). 218 Another phase 2 trial (NCT01282424) of idelalisib treated indolent NHLs including FL, SLL, MZL, and lymphoplasmacytic lymphoma (LPL) with or without WM and showed antitumor activity with an acceptable safety profile (ORR 57% and CR 6%). 219 Moreover, combinations of idelalisib with other novel agents may improve the response rate and DoR. Studies of idelalisib in combination with obinutuzumab in relapsed or refractory FL (NCT03890289) and in combination with BR in indolent B-NHLs and MCL (NCT01088048 and NCT01090414) are ongoing. However, two phase 1 trials of idelalisib and lenalidomide in patients with recurrent FL (NCT01644799) and MCL (NCT01838434) showed emerging toxicities as new combinations. 220 Umbralisib (TGR-1202) and parsaclisib (INCB050465) are also p110δ inhibitors with different chemical structures. 221 A phase 1 trial (NCT02268851) of umbralisib and ibrutinib showed an ORR of 67% (CR 19%) in relapsed or refractory MCL. 222 Duvelisib (IPI-145/INK1197), which is an inhibitor of both p110δ and p110γ, showed efficacy in various types of lymphomas, including DLBCL and MCL, in preclinical studies. 223,224 A phase 2 trial (NCT01882803) of duvelisib monotherapy in relapsed or refractory indolent NHLs demonstrated an ORR of 46% (41% in FL, 33% in MZL, and 68% in SLL). 225 Temsirolimus (CCI-779) is a derivative of rapamycin, and a phase 2 trial (NCT00117598) of temsirolimus as a single agent in relapsed MCL showed an ORR of 38% (CR 3%). 226 In a randomized phase 3 trial (NCT01646021) enrolling patients with relapsed or refractory MCL, significant improvement in PFS and better tolerance were observed in patients treated with ibrutinib vs temsirolimus. 227 Another ongoing study (NCT01653067) is evaluating temsirolimus in combination with DHAP in patients with relapsed or refractory DLBCL. 228 Everolimus (RAD001) is an oral mTOR inhibitor that has been used as a single agent in relapsed or refractory aggressive and indolent NHLs as well as HL. [229][230][231] Clinical trials (NCT03697408 and NCT00967044) to assess everolimus combined with other agents, such as itacitinib and panobinostat, are recruiting patients.

JAK-STAT
The JAK-STAT pathway is activated by extracellular cytokines such as interferons, IL-2, IL-6 and growth factors, which regulate cell survival, proliferation, differentiation, and apoptosis. 232,233 There are four cytoplasmic JAK kinases: JAK1, JAK2, JAK3, and TYK2. JAK1/ JAK3 are prone to immunoregulation, while JAK2 is associated with erythrocyte and platelet formation. 234,235 JAKs lead to STAT phosphorylation, homodimerization, and nuclear translocation (Fig. 1). 233,236 There are seven STAT proteins (STAT1, STAT2, STAT3, STAT4, STAT5A, STAT5B, and STAT6). 234,235 The activation of the JAK/STAT signaling pathway, as assessed by STAT3 or STAT5B phosphorylation, was present in T-NHLs, including anaplastic lymphoma kinase (ALK)-positive and ALK-negative ALCL, 237,238 HTLV-1-associated ATLL, 239,240 and NKTCL. 241,242 Twenty percent of ALK-negative ALCL patients present mutations of the JAK1 and/or STAT3 genes, 237 and approximately 10% of NKTCL patients present STAT3 mutations. 243 Ruxolitinib (INCB018424) is a JAK1/2 inhibitor approved by the FDA to treat myelofibrosis. Ruxolitinib significantly enhanced apoptosis in HL and PMBCL in vitro and promoted survival in a lymphoma xenograft murine model. 244 A phase 2 study (NCT01877005) of ruxolitinib in advanced relapsed or refractory HL showed poor efficacy as monotherapy (ORR 9.4% and CR 0%). 245 Ruxolitinib and navitoclax, a Bcl-2/Bcl-XL inhibitor, reduced the tumor burden and prolonged survival in an ATLL xenograft murine model. 246 A phase 1 study (NCT03681561) of ruxolitinib in combination with nivolumab in relapsed or refractory HL is currently recruiting patients. However, ruxolitinib has off-target effects due to JAK2 inhibition, which may lead to thrombocytopenia, anemia, and neutropenia. 247 Therefore, agents that can selectively inhibit JAK1, such as itacitinib (INCB039110), are expected to better treat lymphomas in view of the risk-benefit ratio. A phase 1/2 study (NCT03697408) of itacitinib in combination with everolimus in relapsed or refractory HL is ongoing. In addition, a phase 1/2 trial (NCT02760485) of itacitinib in combination with ibrutinib in subjects with relapsed or refractory DLBCL is also active. NOTCH NOTCH receptors are single-pass type I transmembrane proteins. Four receptors (NOTCH1-4) are expressed in mammals and share a common structure. Among them, NOTCH1 and NOTCH2 are the most widely expressed receptors and play a role in cell growth, proliferation, survival, and differentiation. 248 NOTCH is cleaved in the transmembrane region by the γ-secretase complex, which can be inhibited by small-molecule γ-secretase inhibitors (GSIs). After release from the membrane, the intracellular portion of the NOTCH receptor translocates to the nucleus, where it interacts with the RBPJ DNA-binding protein and recruits the MAML1 transcriptional coactivator to assemble the transcriptional complex and start transcription. The signal can be terminated by the proteasome (Fig. 1). 249 Mutations of NOTCH1 and NOTCH2 have been reported to mediate the differentiation of B-or T-cell lineages. 250 In T-cell lymphoblastic lymphoma (T-LBL), NOTCH1 mutations vary from 30% to 80%. 251 In DLBCL, NOTCH1 mutations are classified into the N1 subtype, which accounts for 6.1% of ABC DLBCL cases and is associated with poor prognosis. 252 Activation of the NOTCH1 pathway was also observed in MCL, HL and BL. [253][254][255] NOTCH2 mutations are present in approximately 25% of patients with splenic marginal zone lymphoma (SMZL) and approximately 5% of patients with non-splenic MZL 253 and are related to adverse clinical outcomes. 253,256,257 In addition, a similar gene profile has been found in FL. 258 In DLBCL, the BN2 subtype is characterized by BCL6 fusions and NOTCH2 mutations and presents a relatively good prognosis. 252 For targeted agents of the NOTCH pathway, GSIs, as well as antibodies against NOTCH, Delta/Jagged ligands, or other extracellular components involved in the NOTCH signaling cascade, have been tested in multiple clinical trials. 259 GSIs can suppress the release of ICN1 from the membrane and effectively abrogate the activation of NOTCH1 transcriptional programs in cell lines. 260 A phase 1 trial (NCT01363817) evaluating the safety and tolerability of BMS-906024 in subjects with T-LBL was completed. Another study showed strong synergy between glucocorticoids and GSIs. 261 A phase 1/2 trial (NCT02518113) to evaluate LY3039478 in combination with dexamethasone in T-LBL patients was also completed. However, GSIs demonstrated dose-limiting goblet cell hyperplasia of the gut, mainly due to the inhibition of both NOTCH1 and NOTCH2 expression on these tissues. 262 In addition, a phase 1/2 trial (NCT03422679) to investigate the safety, tolerability, and preliminary efficacy of CB-103, a pan-NOTCH inhibitor, is recruiting patients. More research and clinical trials are needed to better understand targeted therapy of the NOTCH pathway.

NF-κB
The NF-κB pathway is one of the key signaling pathways implicated in physiological cellular functions and neoplastic processes. 263,264 Core components of the NF-κB pathway are inhibitors of NF-κB (IκB) proteins, the IκB kinase (IKK) complex, and NF-κB transcription factors, which include RelA/p65, RelB, c-Rel/ Rel, p50, and p52. 265 B-cell associated kinases (BAKs), such as BTK or PI3Kδ, are critical signaling transducers of BCR signaling and can trigger a cascade reaction to form a multiprotein CARD11-BCL-10-MALT1 (CBM) complex. 266 This complex interacts with IKK, the upstream molecule of NF-κB, and promotes NF-κB activation (Fig. 1). [267][268][269][270] The constitutive activation of NF-κB is common in most types of B-NHLs. 269 In DLBCL, NF-κB activity is upregulated in PMBCL and ABC-DLBCL but not in GCB-DLBCL. 173 BCR-dependent NF-κB activation was the highest in the MCD subtype (based on the cooccurrence of MYD88 L265P and CD79B mutations) and BN2 subtype. 252 In CLL, the NF-κB pathway is usually activated through BCR and TLRs. 271 For mucosa-associated lymphoid tissue (MALT) lymphomas, intrinsic BCR activation is associated with an advanced stage.
The NF-κB pathway can be inhibited by directly or indirectly targeting NF-κB components. As a direct targeting agent, pevonedistat (TAK-924/MLN4924), a NEDD8-activating enzyme (NAE) inhibitor, suppresses NF-κB activity by blocking phospho-IκBα degradation. 272 A phase 1 study (NCT03479268) of relapsed or refractory CLL and NHLs is ongoing. HSP90 is a component of the IKK complex and prevents the proteasomal degradation of IKKα and IKKβ. 270 Two phase 1 trials (NCT00647764 and NCT00644072) of the HSP90 inhibitor SNX-5422 in patients with lymphomas were completed. Proteasome UPP is a choreographed system that degrades misfolded proteins in all eukaryotic cells. It plays a role in the processes of cell apoptosis, cell-cycle progression, antigen presentation, and DNA repair. [273][274][275][276] The first step of protein degradation is polyubiquitination, and the proteasome binds the polyubiquitin chain and mediates deubiquitination and then degrades the target proteins to oligopeptides less than 25 amino acids (Fig. 1). 277,278 Inhibition of the pro-survival NF-κB pathway is the main antitumor mechanism of proteasome inhibitors in lymphoma. 279 The targeted drugs and clinical trials related to the proteasome are listed in Table 4. Currently, three proteasome inhibitors (bortezomib, carfilzomib, and ixazomib) are approved for MM or MCL. Bortezomib, a reversible proteasome inhibitor, binds primarily with β5 and, to a lesser extent, with β2 and β1 of the 20S proteasome particle. 280 A phase 2 trial (NCT00063713) of bortezomib in relapsed or refractory MCL reported an ORR of 31% (CR 8%). 281 Another phase 2 trial (NCT00901147) of bortezomib and panobinostat showed an ORR of 43% (CR 22%) in relapsed or refractory PTCL patients. 282 Bortezomib in combination with other agents, such as ibrutinib in MCL (NCT02356458), dexamethasone in CTCL (NCT03487133), and chemotherapeutic regimens, such as gemcitabine, dexamethasone, and cisplatin (GDP) in DLBCL (NCT02542111) and CHOP in T-NHLs (NCT00374699), are currently ongoing. A randomized phase 3 trial (NCT00722137) compared the efficacy of R-CHOP with bortezomib, rituximab, cyclophosphamide, doxorubicin, and prednisone (VR-CAP) in untreated MCL, showing an improved median PFS but increased hematologic toxicity. 283 Moreover, a phase 2 trial of bortezomib, low-dose dexamethasone, and rituximab (NCT00981708) presented an ORR of 85% (CR 3%) in untreated WM. 284 A phase 3 trial (NCT01788020) conducted in WM patients to evaluate bortezomib in combination with dexamethasone, cyclophosphamide, and rituximab is ongoing. Other proteasome inhibitors, including the irreversible carfilzomib and the reversible oral inhibitor ixazomib, have been studied in a variety of clinical trials. Trials of carfilzomib (NCT01336920) alone or in combination with other agents including vorinostat (NCT01276717), romidepsin (NCT03141203), umbralisib (NCT02867618), rituximab (NCT03269552), BR (NCT02187133), R-CHOP (NCT02073097) and R-ICE (NCT01959698) in relapsed or refractory lymphoma are ongoing. Phase 2 trials of ixazomib showed an ORR of 8.3% (CR 0%) in relapsed or refractory FL (NCT01939899) and an ORR of 67% in relapsed or refractory CTCL/PTCL (NCT02158975). Ixazomib in combination with rituximab (NCT02339922) or with ibrutinib (NCT03323151) is currently under evaluation in indolent B-NHLs and MCL. Phase 1/2 trials of ixazomib combined with romidepsin (NCT03547700) in refractory PTCL and with rituximab and lenalidomide as frontline therapy in high-risk indolent B-NHLs (NCT02898259) are ongoing.
Directly targeting signaling pathways and off-target effects remain a major issue of signaling pathway inhibitors. For example, AEs of ibrutinib, such as atrial fibrillation and bleeding-related events, were connected with the irreversible targeting of ibrutinib on BTK signaling in cardiac myocytes and platelets. 285,286 The offtarget inhibition of kinases containing an analogous cysteine residue with BTK C481 may also be crucial to the side effects of ibrutinib. 287 Moreover, drug resistance reduces the clinical efficacy, warranting further investigation on combined treatment and dual inhibitors. BTK C481S in the ibrutinib binding site is associated with ibrutinib resistance 288,289 but can be overcome in combination with venetoclax. 290 mTOR inhibitors show limited long-term effectiveness due to feedback PI3K/AKT activation, while dual PI3K/mTOR inhibitors could be better alternatives.

EPIGENETIC REGULATION AND TARGETED THERAPY
Epigenetic regulation mainly includes DNA methylation, histone acetylation and methylation. Histone acetylation and methylation  regulate the chromatin state. In the active status, chromatin is accessible to transcription factors, which is represented by the enrichment of H3K27 acetylation and H3K4 methylation. In the repressive status, chromatin is compact and inaccessible to transcription factors, which is characterized by the enrichment of H3K36, H3K27 and H3K9 trimethylation (Fig. 2). 291 Epigenetic dysregulation plays an important role in both B-and T-NHLs and represents potential therapeutic targets according to preclinical data and clinical trials.
DNA methylation and targeted therapy DNMT. The main type of DNA methylation observed in mammals is the methylation of CpG dinucleotides. 292 DNA methyltransferases (DNMTs) mediate this process and induce transcriptional repression. DNMT1 maintains DNA methylation on hemimethylated CpG sites, whereas DNMT3A and DNMT3B are involved in DNA methylation on unmethylated CpG sites. In vitro, the molecular silencing of DNMT1 decreased the expression of cellcycle genes, such as CDK1, CCNA2, and E2F2, in GCB-DLBCLderived cell lines. 293 Analysis of DLBCL patients reported the overexpression of DNMT1, DNMT3A, and DNMT3B in 48%, 13%, and 45% of patients, respectively. 294 Moreover, DNMT1 loss induced altered methylation levels and impaired tumor cell proliferation in mice with T-NHLs. 295 Almost all T-NHL subtypes harbor mutations of DNMT3A. 296 The targeted drugs and clinical trials related to epigenetic modifications are listed in Table 5. Azacitidine, a demethylating agent, inhibits DNMTs by incorporating into RNA and DNA through covalent bonding to DNMTs. A phase 1/2 trial (NCT01120834) showed that azacitidine in combination with vorinostat induced an ORR of 6.7% in patients with relapsed or refractory DLBCL. Azacitidine was also studied in combination with R-CHOP in a phase 1/2 trial (NCT01004991) that reported a CR rate of 91.7% in 12 untreated DLBCL patients. In addition, there are some other trials investigating azacitidine plus R-ICE (NCT03450343) or rituximab and GDP (R-GDP) (NCT03719989) in relapsed or refractory DLBCL and azacitidine with CHOP (NCT03542266) in untreated PTCL patients. Decitabine, a DNMT inhibitor, inhibits DNMTs by incorporating into DNA and reversing DNA methylation and transcriptional repression. A phase 1 trial of low-dose decitabine in NHL and CLL reported dose-limiting myelosuppression. 297 Decitabine combined with R-CHOP is being studied in a phase 1/2 trial (NCT02951728) of untreated DLBCL patients with International Prognostic Index (IPI) >1. Moreover, there is a recruiting phase 4 trial (NCT03579082) exploring the efficacy and safety of decitabine, rituximab, with/without DHAP in relapsed or refractory DLBCL. A phase 3 randomized trial (NCT03553537) is comparing the efficacy and safety of decitabine plus CHOP (D-CHOP) vs CHOP alone in patients with untreated PTCL.
Histone methylation and targeted therapy EZH2. Enhancer of zeste homolog 2 (EZH2) functions as a histone methyltransferase and induces transcriptional repression via the trimethylation of H3K27. EZH2 in GC B-cells represses the expression of a set of genes involved in terminal differentiation, such as PRDM1, IRF4, and XBP1, as well as in the negative regulation of cell-cycle progression, such as CDKN1A and CDKN1B. [318][319][320] Mutations of EZH2 occur in 25% of FL and 21.7% of GCB-DLBCL but not in ABC-DLBCL. 319,321 A strong association between EZH2 mutations and the loss of MHC-I or MHC-II expression was found in DLBCL, especially in GCB-DLBCL. 322 A higher level of H3K27me3 at promoters of NLRC5 and CIITA (MHC-I and MHC-II transactivators) was also found in EZH2-mutant cells, 322 indicating the underlying mechanisms of EZH2 mutation on MHC expression. EZH2 mutations also occur in T-NHLs. Tazemetostat, a selective inhibitor of EZH2, can effectively block H3K27 methylation and inhibit mutant lymphoma cells. 323 The phase 1 part of a phase 1/2 study (NCT01897571) of tazemetostat in relapsed or refractory B-NHLs was completed and demonstrated acceptable safety and potential antitumor activity (ORR 38% and CR 14%). 324 Additionally, phase 1 studies are assessing the novel EZH2 inhibitors SHR2554 and PF-06821497 in lymphoma (NCT03603951 and NCT03460977).
KMT2D. Histone-lysine N-methyltransferase 2D (KMT2D), also called MLL2, is a member of the SET1 family of histone methyltransferases and modulates transcription by H3K4 methylation. Integrative genomic analysis identified that KMT2Dtargeted genes included TNFAIP3, TNFRSF14, and SOCS3, which suppress tumorigenesis, and genes involved in cell signaling pathways such as JAK-STAT and BCR. 325 The incidence of inactivating mutations of KMT2D is observed in 72% of FL 319 and 30% of DLBCL. 326 KMT2D missense mutations lead to a significant reduction in H3K4 methylation in vitro. 327 Recent studies in mice showed that the loss of KMT2D resulted in decreased H3K4 methylation and increased tumor development. 325,327 Though there are no targeted agents for KMT2D, the histone deacetylase inhibitors (HDACis) romidepsin and chidamide showed the ability to restore H3K4me3 levels in KMT2D mutant cells in vitro. 328 Chidamide combined with decitabine was observed to induce the apoptosis of Jurkat cells bearing KMT2D mutations in vitro and in vivo. 328 Histone acetylation CREBBP/EP300. The balance between histone acetyltransferases (HATs, including CREBBP and EP300) and HDACs is critical to maintain a normal histone acetylation status in cells. CREBBP and EP300, as histone acetyltransferases, regulate gene transcription by catalyzing the acetylation of the lysine residues of histones. Inactivating mutations in CREBBP and EP300 in GC B-cells decrease p53-mediated tumor suppression and enhance the protooncogenic activity of BCL-6. 329,330 CREBBP mutation is also associated with reduced MHC-II expression, which is a key element in antigen presentation, thereby promoting tumor escape from the immune system. 331 CREBBP and EP300 mutations were found in 65% and 15% of FL, respectively. CREBBP is mutated in DLBCL, with a significantly higher incidence in the GCB subtype (32% in GCB-DLBCL vs. 13% in ABC-DLBCL). Mutations of EP300 were observed in 10% of DLBCL. 329 In PTCL-NOS, CREBBP and EP300 are mutated in 4% and 8% of patients, respectively. 328 In NKTCL, EP300 is mutated in approximately 3.8% of patients. 243 HDACs. HDACs are divided into four groups: HDAC I (HDAC 1, 2, 3, and 8), HDAC II (HDAC 4, 5, 6, 7, 9, and 10), HDAC III and HDAC IV. 332
Although epigenetic alterations show clinical significance, modulators specifically targeting these alterations remain to be developed. Demethylation agents and HDACis have presented clinical efficacy in many lymphoma subtypes. However, the exact mechanisms of action remain unclear, and biomarkers to predict clinical effects need to be further explored. Moreover, monotherapy with epigenetic agents may have limited efficacy in lymphoma in early phase studies. Trials in combination with chemotherapy or other small molecules have demonstrated potent efficacy and acceptable safety and warrant further investigation.

TUMOR MICROENVIRONMENT AND CHECKPOINT-RELATED TARGETED THERAPY
In addition to tumor cells themselves, the tumor microenvironment plays an important role in lymphoma progression. Immunotherapeutic agents can effectively activate the immune system, leading to tumor regression, and have improved clinical outcomes in lymphoma patients. [349][350][351][352] In addition, checkpoint inhibitors combined with CAR-T therapy, epigenetic modulators, radiotherapy, and BTK inhibitors have shown striking efficacy in refractory lymphoma. [353][354][355] PD-1/PD-L1 Programmed cell death-1 (PD-1, also known as CD279) is a member of the immunoglobulin superfamily and functions as an important immune checkpoint that suppresses excessive immune responses. 6 PD-1 is mainly expressed on activated T cells and a small number of B cells, NK cells, activated monocytes, and dendritic cells but is not expressed on naïve T cells. The persistent stimulation of PD-1 on T cells can lead to T-cell exhaustion. 356,357 The ligands of PD-1 include PD-L1 (also known as B7-H1, CD274) and PD-L2 (also known as B7-DC, CD273). 358,359 PD-L1 is expressed on B cells, T cells, dendritic cells, and macrophages. PD-L2 is expressed mainly on dendritic cells, macrophages, mast cells, and certain B cells in response to IL-4 and IFN. 358,359 In addition to those immune cells, PD-L1 is expressed on tumor cells and protects them from immune surveillance; a high level of PD-L1 on tumor cells is associated with poor prognosis in patients. [360][361][362][363] Therefore, PD-1/PD-L1 pathway blockade can promote T-cell activation and cytokine production and preserve the antitumor capacity of T cells in the treatment of lymphomas. 364 PD-1 is overexpressed in the tumor-infiltrating lymphocytes (TILs) of HL, 365 and 94-100% of refractory or relapsed HL cases are positive for PD-L1. 353,366 The 9p24.1 amplification is frequently detected in HL, resulting in increased PD-L1 and PD-L2 expression on Hodgkin and Reed-Sternberg (HRS) cells. 367 Moreover, the amplified 9p24.1 region contains the JAK2 locus, further enhancing PD-L1 expression in HRS cells. 367 In FL, though PD-1 expression on TILs is abundant, PD-L1 expression on lymphoma cells is low (0-5%). [368][369][370][371][372][373][374][375] In DLBCL, the positive rate of PD-1 was 39.5-68.6%, [376][377][378][379][380] and the positive rate of PD-L1 was 24-75%. 375,[380][381][382] Moreover, the number of PD-1 + TILs is higher in the GCB subtype, and patients with PD-L1 + tumor cells have inferior OS compared to those with PD-L1tumor cells. 379 Soluble PD-L1 (sPD-L1), independent of IPI, has been reported to be an adverse prognostic factor for DLBCL. Similar to PD-1, sPD-L1 is elevated in DLBCL patients at diagnosis and returns to normal when patients achieve CR. Thus, sPD-L1 is an effective predictor of DLBCL. 382 In PTCL, PD-1 is positive in 70% and 61% of AITL and PTCL-NOS, respectively, and PD-1 is rarely detected in ALCL. PD-L1 is expressed in 46% of ALK + ALCL and in 46% of ALK -ALCL. In contrast, there is no PD-L1 expression in AITL and PTCL-NOS. 383 The targeted drugs and clinical trials related to PD-1 are shown in Table 6. Nivolumab and pembrolizumab were approved by the FDA to treat relapsed or refractory HL. 366,377,384,385 In a phase 1 trial (NCT01592370) of nivolumab in relapsed or refractory HL, the ORR was 87% (CR 17%), and the 24-week PFS was 86%. 349 In a phase 2 trial (NCT02181738), the efficacy of nivolumab was evaluated in relapsed or refractory HL. At a median follow-up of 8.9 months, the ORR was 66.3% (CR 9%), and the 6-month PFS and OS were 77% and 99%, respectively. 350 In relapsed or refractory NHLs, a phase 1 trial of nivolumab (NCT01592370) showed an ORR of 40% (CR 10%) in FL, 36% (CR 18%) in DLBCL, and 17% (CR 0%) in T-NHLs. 386 Another phase 2 trial (NCT02038933) of nivolumab in ASCT-failed DLBCL showed an ORR of 10.3% (CR 3.4%). Nivolumab in relapsed or refractory ALK + ALCL (NCT03703050) and PTCL (NCT03075553) is currently under clinical evaluation in phase 2 trials. In relapsed or refractory PCNSL and testicular lymphoma (NCT02857426), nivolumab showed an ORR of 100% (CR 80%). 351 In addition, nivolumab combined with BV (NCT02572167) in relapsed or refractory HL had a reported ORR of 82% (CR 16%), 387 and this combination in NHLs (NCT02581631) is ongoing. Nivolumab combinations with other targeted agents such as lenalidomide in relapsed or refractory lymphoma (NCT03015896), rituximab in FL (NCT03245021), cabiralizumab in PTCL (NCT03927105), and in combination with chemotherapy, such as rituximab, gemcitabine, and oxaliplatin (R-GemOx) in elderly lymphoma patients (NCT03366272), R-CHOP (NCT03704714), and rituximab, doseadjusted etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin (DA-R-EPOCH) (NCT03749018) in aggressive NHLs are ongoing.
Pembrolizumab, a humanized mAb of PD-1, showed an ORR of 65% (CR 16%) in a phase 1 trial (KEYNOTE-013, NCT01953692) of relapsed or refractory HL 353 and an ORR of 69% (CR 22.4%) in a phase 2 trial (KEYNOTE-087, NCT02453594). 366 Pembrolizumab induced an ORR of 41% (CR 11%) in transformed DLBCL and showed no response in relapsed or refractory CLL in a phase 2 trial (NCT02332980). 388 In a phase 1 trial (NCT01953692), pembrolizumab was evaluated in relapsed or refractory PMBCL, and the ORR was 41% (CR 11.8%). 389 Trials of pembrolizumab in relapsed or refractory gray-zone lymphoma and PCNSL (NCT03255018) and in untreated B-NHLs (NCT03498612) are ongoing. In a study of relapsed or refractory NKTCL patients who failed asparaginase treatment or ASCT, pembrolizumab presented an ORR of 100% (CR 71.4%). 352 In a phase 2 trial (NCT02243579) of pembrolizumab in advanced relapsed or refractory MF and SS, the ORR was 37.5% (CR 8.3%). Thus, trials of pembrolizumab in MF (NCT03695471) and NKTCL (NCT03728972) are ongoing. Pembrolizumab in combination with other targeted agents, such as umbralisib (NCT03283137), lenalidomide (NCT02875067), mogamulizumab (NCT03309878), rituximab (NCT03401853), pralatrexate (NCT03598998), CAR-T    CTLA-4 CTLA-4, a member of the immunoglobulin family receptors, together with CD28, are homologous receptors of CD4 + and CD8 + T cells. Both receptors share a pair of ligands (CD80/CD86) expressed on the surface of antigen-presenting cells (APCs). In contrast with CD28, the signal of CTLA-4 suppresses the activation of T cells, and the affinity of CTLA-4 and CD80 is higher than that of CD28 and CD80. In addition to APCs, CTLA-4 is also present in resting T cells in the form of intracellular vesicles and expressed on the cell membrane surface when T cells are activated. 390 The CTLA4-CD86 protein recruits and activates Tyk2, leading to STAT3 activation and the expression of genes involved in immune suppression and tumor growth. Although the CTLA-4 antibody ipilimumab 391 has become a first-line therapy in metastatic melanoma, 392,393 the application of the CTLA-4 antibody still needs to be explored in hematological malignancies. 394 The phase 1 part of a phase 1/2 trial (NCT00089076) of ipilimumab in relapsed or refractory B-NHLs induced an ORR of 11.1% (CR 5.6%). 395 Trials of ipilimumab combined with other agents, such as nivolumab (NCT02408861) in HIV-associated HL, are ongoing. A trial of tremelimumab, another CTLA-4 mAb, in combination with durvalumab in relapsed or refractory DLBCL (NCT02549651) was completed.
CD47/SIRPα CD47 is a new immune checkpoint that is expressed in normal cells and upregulated in various tumors. [396][397][398] Its ligand SIRPα is expressed on myeloid cells (monocytes, macrophages, and myeloid dendritic cells). CD47/SIRPα mainly regulates innate immune cell activity and sends out a "do not eat me" signal to escape the attack of innate immune cells. 399 MYC can upregulate the expression of the CD47 gene by binding to the promoter of CD47. The downregulation of MYC gene expression in a murine model led to decreased CD47 expression. 400 CD47 is upregulated in various NHLs (DLBCL, MCL, FL, and CLL) and is associated with poor clinical outcomes in patients. 401 Targeting CD47 can reduce liver and central nervous system metastasis in Raji-engrafted mice, 402 suggesting the association of CD47 with the extranodal metastasis of lymphoma cells. TTI-621 (an anti-CD47 antibody) enhances macrophage-mediated phagocytosis and can effectively control B-NHL growth in xenograft murine models. Another anti-CD47 antibody, Hu5F9-G4, combined with rituximab, is effective in the treatment of NHLs. In the phase 1 part of a phase 1/2 trial (NCT02953509), 22 relapsed or refractory DLBCL and FL patients were enrolled. The ORRs and CR rates were 40% and 33% in DLBCL and 71% and 43% in FL, respectively. The most common AEs were anemia and infusion reactions. 403 OX40/OX40L OX40 is a member of the TNFR superfamily. Under physiological conditions, it is mainly expressed in activated T cells and is more abundant in CD4 + T cells than in CD8 + T cells. OX40L, the ligand of OX40, is a type II transmembrane protein and is expressed in a variety of APCs (B-cells, dendritic cells, and macrophages), activated T cells, vascular endothelial cells and mast cells. [404][405][406] The OX40L-OX40 signaling pathway is the basis for effector T-cell proliferation and memory T-cell development. However, the OX40L-OX40 axis can promote immune escape and tumor growth. Experimentally, an OX40 agonist showed antitumor activity in combination with other drugs. Intratumoral injection of anti-CTLA-4 and OX40 agonists depleted tumor-infiltrating Tregs in murine lymphoma models. 407 A phase 1 clinical study (NCT03636503) of PF-04518600 (the OX40 agonist) in combination with utomilumab (4-1BB agonist) and rituximab or in combination with avelumab (anti-PD-L1) and rituximab is ongoing in aggressive B-NHLs.

Other immune checkpoint molecules T-cell immunoglobulin and ITIM domain (TIGIT) is a coinhibitory receptor that is expressed on NK cells and different types of T cells, including effector and memory T cells and
Tregs. [408][409][410] The ligands of TIGIT, CD155 (PVR) and CD112 (PVRL2, nectin-2) are expressed on APCs, T cells, and tumor cells. 411,412 In NHL, TIGIT and PD-1 are frequently coexpressed on TILs. Approximately 78-83% of CD8 + and 69-70% of CD4 + T effector memory cells (TEMs) are simultaneously positive for these two inhibitory molecules, and these TEMs have limited capability for IL-2, IFN-γ, and TNF-α secretion. 413 In FL, TIGIT is mainly expressed by CD8 + effector and memory T cells and is related to advanced disease stage. 414 TIM-3 inhibits Th1 cell responses, 415 and its antibodies have been found to potently enhance antitumor immunity. 416 An increased number of TIM-3 + T cells is related to the unfavorable prognosis of FL patients. 414 TIM-3 is preferentially expressed on the microvascular endothelial cells of lymphoma, suppresses the activation of CD4 + T lymphocytes and facilitates the progression of lymphoma by mediating immune evasion. 417 Indoleamine 2,3-dioxygenase (IDO), a known immune suppressor, plays a role in human mesenchymal stromal cells (MSCs) to regulate immunity in the tumor microenvironment. IDO + MSCs can inhibit T-cell proliferation in vitro. In a lymphoma murine model, IDO + MSCs could enhance tumor growth, which could be reversed by the IDO inhibitor D-1-methyl-tryptophan (D1-MT). 418 Since MSCs secrete IDO to further suppress T-cell immune responses, umbilical cord-derived MSCs genetically secrete TandAb (a tetravalent bispecific antibody with two CD3 and two CD19 binding sites). In vitro, TandAb can induce the specific lysis of CD19 + cell lines in the presence of T cells, and an IDO inhibitor could enhance the cytotoxicity of T cells triggered by MSC-TandAb. 419 Clinical studies of IDO inhibitors in lymphomas are still lacking. V-domain immunoglobulin suppressor of T-cell activation (VISTA) is another checkpoint molecule that has a strong inhibitory influence on T cells. 420 VISTA is constitutively expressed in CD11b high myeloid cells and is expressed at a low level on T cells and Foxp3 + CD4 + Treg cells. 421 In animal models with solid tumors, myeloid-derived suppressor cells (MDSCs) infiltrating tumors were found to highly express VISTA compared to peripheral blood cells. 422,423 In a murine model of squamous cancer, anti-VISTA monotherapy increased the infiltration and activation of T cells. A clinical trial (NCT02812875) evaluating the efficacy and safety of CA-107 (targeting PD-L1, PD-L2, and VISTA) for the treatment of lymphoma is ongoing.

ADOPTIVE T/NK-CELL THERAPY
Adoptive T-cell transfer is an emerging immunotherapy in a variety of tumors, particularly CAR-T therapy. In 2017, the FDA approved tisagenlecleucel (a CD19-specific 4-1BB-CAR construct) for the treatment of relapsed or refractory B-ALL, and in 2018, the FDA approved axicabtagene ciloleucel (a CD19-specific CD28-CAR construct) for the treatment of relapsed or refractory DLBCL. Another CD19 CAR-T cell line, lisocabtagene maraleucel (a CD19-specific 4-1BB-CAR construct), is also undergoing evaluation.

CAR-T therapy in lymphoma
In a single-arm, multicenter clinical trial (NCT02348216) for relapsed or refractory DLBCL, transformed FL, and PMBCL, axicabtagene ciloleucel had an ORR of 83% (CR 58%) and median PFS of 5.9 months. 424 In another clinical trial (NCT02030834) of relapsed or refractory B-NHLs, tisagenlecleucel induced an ORR of 64.3% (CR 57.1%). Moreover, all CR patients were still in remission at 6 months. 425 In a phase 2 trial (NCT02445248) of tisagenlecleucel in relapsed or refractory DLBCL, the ORR was 52% (CR 40%), with a 1-year RFS of 65%. 426 In addition to axicabtagene ciloleucel and tisagenlecleucel, lisocabtagene maraleucel was tested in relapsed or refractory DLBCL, PMBCL, FL, and MCL (TRANSCEND, NCT02631044) and showed a CR rate of 80% in high-grade B-cell lymphoma (double/ triple hit) and DLBCL. 427,428 A phase 1 dose-escalation study (NCT03355859) of anti-CD19 JWCAR029 was conducted in refractory B-NHLs, and the ORR was 100%, with 6 of 9 (66.7%) evaluable patients achieving CR. In this study, core needle biopsy was performed on tumor samples on day 11 after CAR-T cell infusion. Further RNA sequencing of these tumor samples identified gene expression signatures differentially enriched in complete and partial remission patients. Increased tumorassociated macrophage infiltration was negatively associated with remission status. 429 In addition to studies targeting CD19 CAR-T cells, studies on CD20, CD22, and CD30 CAR-T cell therapy have also been carried out. In a phase 2 study (NCT01735604) of anti-CD20 CAR-T therapy, the ORR was 81.8% (CR 54.5%). 430 In a phase 1 trial (NCT02315612), anti-CD22 CAR-T cells were evaluated in patients with B-cell malignancies resistant to CD19 CAR-T cells and showed a CR rate of 73%, with a median remission duration of 6 months. 431 A phase 1 trial (NCT01306146) of anti-CD30 CAR-T cells showed a CR rate of 28.6% in relapsed HL and a CR rate of 50% in ALCL. 432 Anti-CD4 CAR-T cells could control the growth of tumors in a xenograft murine model of ALCL. 433 However, this therapy also faces the challenge of CAR-T cells sharing antigens with normal T cells and can recognize and kill three types of cells: tumor T cells, normal T cells, and CAR-T cells. This problem can lead to the "autophase killing" of CAR-T cells, while CAR-T cells targeting normal T cells may lead to severe infection in patients. 434 Therefore, reducing the side effects of CAR-T cells in T-NHLs has become the focus of research. Moreover, using CRISPR/Cas9 gene-editing technology, generating CAR-T cells (also known as UCART7) that lack CD7 and TCR alpha-chain expression could target CD7 + T-cell malignancies and reduce mutual attacks between CAR-T cells. 435 Although CAR-T therapy has been successful in the treatment of hematological malignancies, there are still patients who do not respond to the treatment, as well as some patients presenting signs of AEs such as severe cytokine release syndrome (CRS), infection, and neurotoxicity. Therefore, identifying patients who may respond to CAR-T therapy and patients who may have serious side effects during treatment has become a research hotspot.

CAR-NK therapy
With the continuous development of CAR-T therapy, CAR-NK cells have also become a focus of attention. NK cells are cytotoxic immune cells that form a small fraction of normal lymphocytes and can trigger the innate immune response against tumor cells and virus-infected cells. 436 Studies have shown that NK cells have a nonnegligible role in tumor monitoring, and loss of NK cells leads to tumor progression. [437][438][439] Because NK and T cells are functionally similar, NK cells can also be used to attack tumors. Many researchers hope that CAR-NK cells can achieve results in tumor treatment similar to CAR-T cells. Compared with T cells, NK cells kill tumor cells in a nonantigen-dependent manner. Moreover, NK cells express CD56 and CD7 but lack the expression of CD3, TCR, and CD5. 440 When used in the treatment of T-NHLs, the fratricide of CAR-NK cells was reduced. 441 Anti-CD19 cord blood (CB)-derived NK cells were evaluated in a xenograft lymphoma murine model and significantly prolonged the survival of mice. 442 In another study, anti-CD5 CAR-NKs had potent antitumor activity against a variety of T-NHLs and primary tumor cells in vitro and in a murine model. 443 Clinical trials (NCT03383965, NCT029170083 and NCT03049449) of CAR-NK cells targeting CD19, CD20, and CD22 have begun for the treatment of B-NHLs. In addition, a phase 1 trial (NCT02742727) of anti-CD7 CAR-NK for the treatment of T-cell malignancies is ongoing.

SPECIFIC ONCOGENES AND PROTEINS RELATED TO TARGETED THERAPY
Specific oncogenes, such as MYC, BCL-2, and BCL-6, converge proliferation, differentiation, and anti-apoptotic signaling in lymphoma cells and play critical roles in lymphomas. Moreover, lymphomas that have a concomitant translocation of MYC and BCL-2 or BCL-6 represent high-grade B-cell lymphoma and are resistant to conventional R-CHOP chemotherapy. The tumor suppressor gene p53 is involved in the process of DNA repair, and the depletion or mutation of p53 promotes lymphoma progression and drug resistance. The t(2;5)(p23;q35) translocation results in the NPM1/ ALK fusion protein and then activates the downstream oncogenic transcription factor STAT3, enhancing lymphoma cell proliferation and growth. Thus, these specific oncogenes are greatly involved in lymphoma genesis and progression, and targeting these genes and their downstream pathways might retard tumor progression and improve patient survival.

MYC
MYC is a family of three proto-oncogenes that function as important regulators of cell proliferation, growth, differentiation, and apoptosis. They encode the related transcription factors MYC, MYCN, and MYCL, also known as c-MYC, N-MYC, and L-MYC, respectively. 444 The Ig-MYC translocation is the most common type of MYC alteration and can cause MYC overexpression. 445 MYC is expressed at the pro-B and pre-B-cell stages and in a minority of GC B-cells. [446][447][448] Additionally, MYC is frequently overexpressed in lymphomas of GC origin. In BL, the t(8;14) translocation is found in approximately 80% of all patients. 449 In DLBCL, MYC overexpression is shown in 30-50% of patients. 450,451 MYC translocations preferentially occur in GCB-DLBCL over ABC-DLBCL (17.7% vs. 6.7%). 450 A high level of MYC is associated with a low treatment response and poor prognosis in DLBCL patients treated with R-CHOP and may also lead to an increased relapse rate in the central nervous system. 452,453 Studies have shown the potential effect of MYC-associated agents, including targeting cell-cycle-associated vulnerabilities, transcription, RNA processing and turnover, ribosome biogenesis and translation, as well as MYC-induced metabolic perturbations. 454 The mitotic spindle-regulatory kinases Aurora-A and Aurora-B are both overexpressed in MYC-associated B-cells, and Aurora-A promotes the stabilization of MYC and MYCN. 455,456 The targeted drugs and clinical trials related to specific oncogenes and proteins are shown in Table 7. A phase 1 trial (NCT01897012) of alisertib combined with romidepsin in relapsed or refractory NHLs is ongoing.

BCL-2
The BCL-2 family of proteins regulates the intrinsic pathway of mitochondrial apoptosis 457 and can be divided into three groups: anti-apoptotic proteins (BH1-4 domains), multi-BH domain proapoptotic proteins (BH1-3 domains), and BH3-only pro-apoptotic proteins. The t(14;18)(q32;q21) translocation is a common type of BCL-2 translocation. 458 Mutated BCL-2 affects cells in several aspects, such as proliferation, apoptosis, angiogenesis, and metastasis, resulting in the development of hematological malignancies. 459,460 BCL-2 translocation is the major hallmark of FL (>80% of samples); it occurs in bone marrow pre-B cells and leads to high BCL-2 protein expression. 461 Chromosome 18q21 amplification leads to BCL-2 overexpression and is observed in patients with MCL. 462 BCL-2 overexpression is also detected in approximately 30% of DLBCL. 463 The term double-hit lymphoma (DHL) refers to a subset of DLBCLs that present concurrent rearrangements of MYC and BCL-2 (sometimes BCL-6). 464 DHL is present in 5-10% of DLBCL and is mostly classified as the GCB subtype, with highly aggressive clinical behavior and poor response to frontline regimens. 465,466 The term double-expressor lymphoma (DEL) refers to a subset of DLBCLs that show the coexpression of MYC (>40%) and BCL2 (>50%) by immunohistochemistry in the absence of chromosomal translocations. DEL is present in 25-30% of DLBCL and is mostly classified as the ABC subtype, which is also associated with poor clinical outcomes. 466,467 ABT-737, which binds to BCL-2, BCL-XL, and BCL-w with high affinity, had promising preclinical effects in CLL. 468,469 Navitoclax (ABT-263), the orally available derivative of ABT-737, 470 was shown to provoke transient thrombocytopenia in phase 2 trials of patients with B-NHLs due to the importance of BCL-XL for the survival of platelets. 471 A phase 1 trial of navitoclax combined with rituximab (NCT00788684) in lymphoid cancers is ongoing. Venetoclax (ABT-199), a highly selective BH3 mimetic, is designed to treat lymphomas with BCL-2 translocations. A phase 1 trial (NCT01328626) of venetoclax in relapsed or refractory NHLs showed an ORR of 75% (CR 21%) in MCL, an ORR of 38% (CR 14%) in FL, an ORR of 18% (CR 12%) in DLBCL and an ORR of 67% (CR 0%) in MCL. 472 A phase 2 study (NCT02187861) of venetoclax plus rituximab vs. venetoclax plus BR in patients with relapsed or refractory FL was completed. The results showed an ORR of 32.7% (CR 13.2%) in the venetoclax plus rituximab group, an ORR of 45.1% (CR 27.5%) in the venetoclax plus BR group, and an ORR of 51% (CR 23.5%) in the BR group. Many clinical trials on combination therapy of venetoclax and chemotherapy or other targeted agents are active. In MCL, a phase 3 randomized, double-blind study (NCT03112174) to compare the efficacy and safety of the combination of ibrutinib and venetoclax vs. ibrutinib and placebo is ongoing. A phase 1 study (NCT03255096) on the combination of RO6870810 (a bromodomain inhibitor) and venetoclax, with or without rituximab, in relapsed or refractory DLBCL and high-grade B-cell lymphoma is ongoing. To test the effect of venetoclax in combination with chemotherapy, a study (NCT02055820) of venetoclax in combination with R-CHOP or obinutuzumab plus CHOP (G-CHOP) in previously untreated DLBCL was performed, and the results demonstrated an ORR of 87.5% (CR 79.2%) in the venetoclax plus R-CHOP group and an ORR of 87.5% (CR 78.1%) in the venetoclax plus G-CHOP group. Moreover, 87.5% of DEL patients achieved CR. 473 Another phase 1 trial (NCT03036904) of venetoclax plus DA-R-EPOCH is also active for aggressive B-NHLs. BCL-6 BCL-6 was initially discovered as an oncogene in B-NHLs. The BCL6 protein is an evolutionarily conserved zinc finger transcription factor with an N-terminal broad-complex, tram track and bric-a-brac/Pox virus and zinc finger (BTB/POZ) domain and functions as a transcriptional repressor. 474 Transcription factors, transcriptional corepressors, signaling mediators, and catalytic enzymes can be regulated by BCL-6. Studies have shown that BCL-6 overexpression inhibits reactive oxygen species (ROS) generation and represses the apoptosis induced by chemotherapy in B-NHL cells. 475,476 Similar to BCL-2, BCL-6 is the key factor for the development and maintenance of GCs within lymphoid follicles. Once GC B-cells begin their differentiation into memory B-cells and PCs with an appropriate affinity for the inciting antigen, BCL-6 will be phosphorylated and subsequently degraded by the proteasome. 476 Moreover, BCL-6 regulates T FH cell differentiation. 477,478 BCL-6 translocations are found in 40% of DLBCL, 48% of nodular lymphocyte-predominant Hodgkin lymphoma, and 5-10% of FL. 445,479,480 ABC-DLBCL patients have more BCL6 translocations than GCB-DLBCL patients (24% vs. 10%). In T-NHLs, BCL-6 is detectable in some types of PTCL, especially ALK + ALCL and lymphomas derived from T FH cells, particularly AITL. 481,482 Oncogene addiction is switched to BCL-2 and BCL-XL in the context of BCL-6 inhibition. 483 To solve this problem, a combined treatment of RI-BPI (a BCL-6 inhibitor) and ABT-737 might be a choice but needs more experimental verification.

p53
The p53 transcription factor plays an important role in regulating cell survival by activating gene transcription that is involved in apoptosis and other biological functions. 484 Notably, p53 can interact with the BCL-2 pathway by directly and indirectly regulating the anti-apoptotic activity of the BCL-2 family of proteins. 485 With a negative feedback response, the E3 ubiquitin ligase MDM2 can bind p53 for degradation, maintaining a low expression level of p53 under normal conditions. 486 The dysregulation of p53 can be found in many types of lymphomas, including DLBCL (16-30%), MCL (21-45%), FL (9-29%), and MZL (8-12%). 458 It is often regarded as an independent prognostic factor for poor outcomes and a signal for chemotherapy resistance. 487 Targeting p53 can potentially restart apoptosis and trigger cell death. Idasanutlin (RG7388), a potent and selective MDM2 antagonist, when combined with obinutuzumab and venetoclax, showed significant antitumor activity in xenograft models. 488 A phase 1/2 trial (NCT03135262) of idasanutlin in combination with rituximab and venetoclax in relapsed or refractory DLBCL patients is ongoing. Selinexor, an inhibitor of exportin 1 (XPO1), inhibits the nuclear export of p53 and restores p53 nuclear localization. A phase 1 study (NCT01607892) of selinexor showed an ORR of 31% (CR 6%) in advanced NHLs. 489 A study of selinexor combined with chemotherapy (NCT03147885) in advanced B-NHLs is ongoing.
ALK ALK + ALCL is characterized by the expression of ALK fusion proteins. 464 The major type of ALK fusion is the t(2;5)(p23;q35) translocation, which is detectable in approximately 75-85% of ALK + ALCL. 490,491 All fusion proteins can activate the downstream oncogenic transcription factor STAT3 and promote proliferation and growth in cancer cells. 492 Many inhibitors are clinically available for targeting ALK tyrosine kinase activity, including crizotinib as a first-generation agent, ceritinib and brigatinib as second-generation agents and lorlatinib as a thirdgeneration agent. 493 The study of crizotinib in relapsed and refractory ALK + lymphomas showed an ORR of 90.5%, with a 2year PFS of 63.7% and a 2-year OS of 72.7%. 494 Thus, a phase 2 trial of crizotinib (NCT02419287) and brigatinib (NCT03719898) in relapsed ALK + lymphomas is ongoing. Moreover, because of acquired resistance from firstand second-generation agents, a phase 2 study (NCT03505554) to define the ORR of lorlatinib in patients with ALK + lymphomas resistant or refractory to ALK inhibitors is ongoing.

CONCLUSIONS
With the understanding of the biological function of surface markers, signaling transduction pathways, and epigenetic modulations as well as the orchestration of the microenvironment with lymphoma cells in lymphoma progression, many novel agents and immune therapeutic strategies have been developed. These therapies enable clinicians to perform precision medicine and significantly improve the prognosis of patients. However, many questions remain to be answered, such as treatment scheduling, optimized dosage and combinations with other agents. The identification of potential biomarkers that can predict the clinical responses and toxicities of these targeted therapies is challenging. In conclusion, mechanism-based targeted therapy is a promising strategy to eventually make lymphoma a curable disease.